Choosing between Fault-Tolerance and Increased V&V for Improving Reliability



The paper by Peter Popov, Lorenzo Strigini and Bev Littlewood was presented at the International Conference on Parallel and Distributed Processing Techniques and Applications (PDPTA'2000), June 26-29, 2000, Monte Carlo Resort, Las Vegas, Nevada, USA), pp. 535-540, ISBN 1-892512-22-x, Copyright CSREA Press. The full text is available as a .pdf file. The abstract is given below.


Abstract
Fault tolerant systems based on the use of software design diversity may be able to achieve high levels of reliability more cost-effectively than other approaches, such as heroic debugging. Earlier experiments have shown that multi-version software systems are more reliable than the individual versions. However, it is also clear that the reliability benefits are much worse than would be suggested by naive assumptions of failure independence between the versions.
To decide whether to use design diversity or other means for achieving the desired reliability  a developer would need to know how they compare from the viewpoint of cost-effectiveness. Empirical data are insufficient for deciding this question, and expert opinions differ. We refute a recently published argument in favour of diversity and in the process show some general factors deciding whether process improvement, or debugging of the versions in a multiple-version system, will increase or decrease the statistical correlation between  failures of the versions. The conclusion is that there is as yet no evidence that the choice between design diversity and other means of reliability improvement can be decided by general arguments rather than by detailed (and uncertain) special-case analysis.

N-version design Versus one Good Version



The article by Bev Littlewood, Peter Popov and Lorenzo Strigini was presented at the International Conference on Dependable Systems & Networks (FTCS-30, DCCA-8), 24-27 June, 2000, New Your City, (USA). The text appears in the 'Workshops and Fast abstracts' Digest of DSN'2000, pp. B42-B43. The text is available for download as a .pdf file.
Page maintained by: Peter Popov
Last modified 31 January 2000.